This paper looks at data Lakehouse architectures as a game changer in enterprise data infrastructure, focusing on Apache Iceberg storage. We cover the full capabilities of these systems for data throughout its life cycle – from ingest to visualization—and how machine learning can be used to enhance that. We also look at execution frameworks based on directed acyclic graphs and the privacy implications of those workflows. Our results show this integrated approach is better for operational efficiency, analytical flexibility, and compliance than traditional, siloed architectures.

Flying Ad hoc Networks (FANETs) are formed by unmanned aerial vehicles (UAVs) that communicate over wireless links without fixed infrastructure. High node mobility, frequent topology changes and the three-dimensional nature of the environment make routing a challenging task. This paper presents a compact systematic review of routing strategies in FANETs. We classify existing protocols into five main families: topological (proactive, reactive and hybrid), geographic (position-based), cluster-based (hierarchical), opportunistic (delay-tolerant) and intelligent (AI-based and bio-inspired). Typical performance indicators such as packet delivery ratio, end-to-end delay, throughput, routing overhead and energy consumption are summarized, and their relationship with different protocol classes is discussed. The analysis shows that most existing solutions are adaptations of MANET/VANET algorithms and usually optimize only a subset of metrics. Key research gaps include energy-aware routing, stable operation under extreme mobility and sparse topologies, and deeper integration of learning-based methods. The review provides a concise overview of current approaches and outlines directions for future research in FANET routing.

In the era of cloud-native computing, achieving high scalability, resilience, and performance has become a fundamental requirement for modern application development. This paper presents a critical review of Spring Boot, a leading Java-based framework, and its role in elevating application performance within distributed and microservices-oriented architectures. The study examines Spring Boot’s core features—such as embedded servers, auto-configuration, actuator endpoints, and integration with containerization and orchestration tools like Docker and Kubernetes—that streamline deployment and operational efficiency. Furthermore, it evaluates performance optimization techniques, fault-tolerance mechanisms, and scalability patterns enabled by Spring Cloud and reactive programming models. Through comparative analysis and case-based discussion, the review highlights both the strengths and limitations of Spring Boot in building resilient, cloud-native systems. The findings underscore Spring Boot’s effectiveness in simplifying complex infrastructure concerns while ensuring agility, observability, and robustness in modern software ecosystems.

This article examines the transformative role of artificial intelligence in enterprise cybersecurity monitoring, addressing the fundamental challenges that traditional security operations centers face in managing the exponentially growing volume of security events across complex digital environments. The article explores how machine learning approaches for anomaly detection enable organizations to identify threats without explicit programming for each variant, while also addressing the critical problem of alert fatigue through intelligent prioritization and correlation mechanisms. The article analyzes emerging human-AI collaboration models that redefine security workflows and distribute cognitive load optimally between analysts and automated systems, emphasizing the importance of explainable AI for building appropriate trust. Finally, the article examines future directions toward autonomous security response, identifying current limitations and promising approaches for safe partial-automation while considering regulatory frameworks and adversarial adaptation. Throughout the analysis, the article demonstrates how AI integration represents not merely a technological evolution but a strategic necessity for maintaining viable security operations in an increasingly complex threat landscape.

The convergence of DevOps practices with artificial intelligence (AI) and cloud-native architectures represents a transformative evolution in contemporary software engineering. Organizations increasingly require frameworks that harmonize agility, reliability, and security while managing the complexities of distributed systems, microservices, and continuous delivery pipelines. This research explores the integration of AI-driven mechanisms within DevOps to automate vulnerability management, patch deployment, and demand forecasting, thereby optimizing operational efficiency and reducing exposure to cyber threats. Through an exhaustive review of seminal literature and empirical studies, the paper identifies key dimensions of DevOps, including culture, automation, measurement, and sharing, and investigates how AI interventions can augment these dimensions to enable predictive, real-time security management. The methodology synthesizes theoretical and practical insights from canonical texts, cloud deployment frameworks, microservices observability tools, and AI-driven security systems to propose a comprehensive conceptual model for secure, automated, and intelligent CI/CD pipelines. Findings indicate that AI-enhanced DevOps frameworks significantly improve patching efficiency, reduce system downtime, facilitate intelligent orchestration of resources, and enhance the overall security posture of cloud-native applications without impeding delivery velocity. This research contributes both theoretically and practically by delineating pathways for integrating AI within DevOps pipelines, highlighting operational limitations, and proposing future directions, including adaptive orchestration, standardized observability protocols, and governance models for hybrid and containerized cloud environments.

The rapid proliferation of mobile applications, IoT-enabled systems, and complex multi-vendor infrastructures has intensified the challenges of ensuring system security, stability, and operational integrity. Traditional manual security assessments are increasingly inadequate to address the volume, diversity, and dynamic nature of contemporary software and hardware ecosystems. This study examines the integration of automated vulnerability assessment and penetration testing (VAPT) with advanced machine learning models, threat intelligence, and fuzzy logic-based control strategies to enhance the detection and mitigation of security risks while ensuring system stability. The research synthesizes methodologies from static and dynamic code analysis for Android applications, automated penetration testing in multi-vendor systems, and stability analysis of power electronic converters using state-space techniques. Additionally, the work explores the application of Internet of Things (IoT) frameworks in monitoring critical infrastructure, agricultural systems, and mobile devices, highlighting the importance of real-time threat intelligence and adaptive detection mechanisms. Key contributions include a detailed evaluation of AI-enhanced penetration testing frameworks, a theoretical model for fuzzy logic-based control in series-parallel resonant converters, and a comprehensive discussion on integrating continuous security testing into DevSecOps pipelines. The findings suggest that combining automated VAPT, predictive machine learning models, and advanced control theory can significantly improve detection accuracy, reduce false positives, and enhance overall system resilience. The study provides a multidisciplinary perspective, emphasizing both cybersecurity and system stability considerations, offering practical guidance for researchers and practitioners in deploying robust and intelligent monitoring frameworks.

This article traces the evolution of escape rooms through the case of the Titanic Escape Room, framed as a cyber-physical system in which architecture, firmware, and hardware solutions are subordinated to the task of creating an invisible fail-safe technological substrate for an immersive experience. The study is relevant because the Location-Based Entertainment (LBE) industry is developing rapidly, determining success not through isolated puzzles but by sustaining a cinematic, plausible, and uninterrupted narrative. The article demonstrates that an experience-centric model replaces puzzle-centric design in this context. Through the application of cyber-physical systems theory and HCI principles, combined with fault-tolerant design, this work is novel due to its thorough engineering analysis of the Titanic system. The article proposes a distributed three-tier architecture (COGS server, Raspberry Pi media servers, Arduino nodes). This approach justifies its choice, also justifying the selection of communication protocols and failover strategies. Also examined in detail is the article's analysis of three engineering solutions: adaptive Morse-code decoder, hardware printer redundancy, and hybrid PCM/MP3 audio playback. These are interpreted as manifestations of a unified strategy of engineering for time and reliability, aimed at minimizing latency and covertly eliminating faults, which are preconditions for preserving immersion. The main findings underscore that immersiveness in LBE systems is a reproducible outcome achieved by strict adherence to real-time principles, adaptive interface design, and multilayer redundancy. The article will be helpful to researchers of cyber-physical systems, entertainment-technology engineers, HCI specialists, and developers of immersive LBE projects.

The advent of Large Language Models (LLMs) has fundamentally altered the cybersecurity landscape, specifically within the domain of social engineering. While LLMs facilitate productivity, they also empower threat actors to generate hyper-personalized, grammatically perfect, and contextually relevant phishing campaigns at scale. This paper explores the intersection of generative AI, cognitive psychology, and intrusion detection to propose a novel defense framework. We investigate the efficacy of current AI-driven social engineering tactics, utilizing the Five-Factor Model of personality to map cognitive vulnerabilities exploited by generative agents. Furthermore, we introduce a Context-Aware Defense System (CADS) that leverages fine-tuned LLMs to detect semantic anomalies and psychological manipulation triggers in real-time communications. Our methodology involves simulating high-fidelity spear-phishing attacks against generative agent personas representing diverse psychological profiles. Results indicate that traditional signature-based detection fails against LLM-generated content, whereas the proposed semantic analysis approach improves detection rates significantly. We find that high Agreeableness and Neuroticism correlate with higher susceptibility to AI-generated pretexts. The study concludes that effective defense against the next generation of social engineering requires a paradigm shift from static filtering to dynamic, psychological, and semantic content analysis.

Background: Modern supply chains—especially in healthcare and agriculture—face mounting pressures from sustainability imperatives, technological disruption, public-health requirements, and complex social responsibilities (Duque-Uribe et al., 2019; Haji et al., 2022). Emerging technologies such as blockchain and artificial intelligence (AI) offer transformative potential, yet their adoption raises novel operational, ethical, and governance questions (Dutta et al., 2020; Elufioye et al., 2024).

 Aim: This article develops a publication-ready, theory-informed, and practice-oriented integrative framework that synthesizes sustainable supply chain management (SSCM) principles with blockchain-enabled transparency and AI-driven predictive analytics to improve environmental, social, and operational performance in healthcare and agricultural supply chains (Mani et al., 2016; Elabed et al., 2019).

 Methods: Using a rigorous, narrative-synthesis approach anchored in the provided literature, the study constructs conceptual linkages across sustainability dimensions, technology affordances, stakeholder partnerships, and measurement approaches. The method entails deep theory elaboration, cross-referencing of empirical and conceptual studies, and development of testable propositions and managerial pathways (Hsu et al., 2013; Govindan et al., 2015).

 Results: The resulting framework identifies four integrative pillars—Governance & Partnerships; Technological Transparency (Blockchain); Predictive & Prescriptive Analytics (AI); and Social-Environmental Performance Management—each mapped to specific mechanisms, barriers, and evaluation metrics. The framework explicates how blockchain can reduce waste and increase traceability (Chowdhury, 2025; Dutta et al., 2020), how AI can optimize demand forecasting and reduce resource inefficiencies (Elufioye et al., 2024), and how combined solutions support resilience against disruption while advancing social sustainability and patient safety (Grumiller et al., 2022; Kanokphanvanich et al., 2023).

Originality: This review offers a holistic synthesis, bridging the gap between ULP design methodology and its formal verification requirements, providing a foundational resource for researchers and practitioners navigating the dual constraints of energy efficiency and functional integrity in next-generation VLSI.

Background: Suicide remains a leading public health concern internationally, with measurable changes over recent years that highlight both progress and persistent vulnerabilities in mental health systems (National Institute of Mental Health, 2024; Saunders & Panchal, 2023). Concurrently, development and deployment of large language models (LLMs) and AI-augmented mental health applications are accelerating, producing a contested landscape of opportunity and risk for suicide prevention and mental healthcare broadly (Omar et al., 2024; Karabacak & Margetis, 2023).
Objective: This article synthesizes extant literature to construct a comprehensive, publication-ready research manuscript that: (1) examines how LLMs encode clinical knowledge and their potential utility for mental healthcare and suicide prevention (Singhal et al., 2023; Omar et al., 2024); (2) situates LLMs within persistent structural barriers to access and delivery of mental health services (Ziller, Anderson & Coburn, 2010; Donohue, Goetz & Song, 2024; Coombs et al., 2021); (3) articulates principal technical risks (hallucinations, dataset quality, domain drift) and governance challenges (Islam et al., 2025; Chen et al., 2024; Wettig et al., 2024); and (4) proposes a detailed, ethically grounded methodological framework for evaluation, validation, and staged integration of LLM-based tools into clinical and community settings.
Methods: We performed an integrative synthesis of the provided sources, mapping empirical evidence on suicide epidemiology and service access to contemporary technical literature on LLM capabilities, training-data concerns, and evaluation strategies. From this synthesis we derived a multi-modal research framework combining qualitative stakeholder inquiry, simulated and retrospective validation experiments, prospective safety trials, and continuous monitoring guided by hybrid human-AI oversight. Each element is detailed with operational procedures, measurement constructs, and ethical safeguards drawn from the literature.
Results: The synthesis reveals convergent themes: (1) suicide prevention needs precise, equitable, and accessible interventions; (2) LLMs exhibit surprising clinical pattern understanding but retain unpredictable failure modes and hallucinations; (3) disparities in access to care create both need and risk when AI systems are unevenly distributed or poorly validated in underserved populations; (4) robust evaluation requires domain-specific high-quality data, multi-language and demographic validation, human-feedback loops, and transparency metrics.
Conclusions: LLMs can augment suicide prevention and mental healthcare, but safe, equitable deployment requires methodical evaluation, domain-specific fine-tuning with quality-controlled data, human-in-the-loop safeguards, and policy frameworks to mitigate access-related harms. The proposed research framework operationalizes these requirements and outlines steps for translational research aimed at realizing benefits while minimizing risks.

This article explores neuroarchitecture, as it represents an interdisciplinary approach to designing educational spaces that integrates advances in neuroscience, cognitive psychology, and architectural practice. Addressing of two contemporary challenges is what conditions the study’s relevance: the rise of mental-health problems among all students and the mismatch between customary school environments as well as the pedagogical demands of the twenty-first century. This study aims to identify and systematize principles of neuroarchitecture that enhance engagement, mitigate stress, and support cognitive productivity. The article is novel since it synthesizes theoretical foundations with evidence-based design solutions. After that, the article verifies them through the Shanti Elementary School project case study in the United States. The chief findings do indicate that biophilic designs and optimized light with color plus spatial flexibility not only restore attention but also reduce anxiety, and gain academic achievement, reduce absenteeism, plus heighten resilience among teaching staff. The paper emphasizes that the educational environment significantly influences both cognitive and emotional experiences, and that policy and practice should consider integrating neuroarchitectural strategies as a quality standard in education. Architects and educators in cognitive psychology and neuroscience, as well as researchers and academic administrators, will find the article to be useful.

The increasing sophistication of financial fraud in the U.S. banking system requires advanced and transparent detection mechanisms. This study proposes a blockchain-enabled machine learning framework that enhances fraud detection accuracy and data integrity. Using an open-source dataset from the UCI Machine Learning Repository, five supervised models—Logistic Regression, Decision Tree, Random Forest, Gradient Boosting, and Neural Network—were trained and evaluated. Data preprocessing included feature scaling, encoding, and class balancing to ensure reliability and model generalization. Results show that integrating blockchain’s immutable ledger with artificial intelligence significantly improves detection performance. The Neural Network model achieved the best results with 99.1% accuracy, 98.6% precision, 98.9% recall, and a 98.7% F1-score, outperforming all other algorithms. The blockchain layer ensured data transparency, traceability, and tamper resistance throughout the detection process. This research demonstrates that combining blockchain and AI can strengthen fraud prevention, enhance regulatory compliance under U.S. financial laws, and foster greater trust in digital banking operations. The proposed system offers a scalable and secure foundation for the next generation of fraud detection in the U.S. financial sector.

Public road construction projects represent one of the most financially significant and institutionally complex categories of public procurement worldwide. Persistent challenges such as cost escalation, time overruns, inflation exposure, opportunistic renegotiations, and suboptimal contractor performance continue to undermine value for money, fiscal sustainability, and public trust. Drawing strictly on foundational and contemporary literature in procurement theory, contract economics, and construction management, this study develops an integrated analytical framework to explain why cost overruns and inefficiencies persist despite decades of reform. By synthesizing incentive theory, transaction cost economics, political economy, and empirical findings from infrastructure projects across multiple jurisdictions, the article examines how inflation dynamics, contract incompleteness, governance structures, judicial efficiency, and strategic behavior interact within road construction procurement systems. Special attention is given to the tension between rigid rule-based procurement and flexible relational contracting, as well as the unintended consequences of competitive bidding under asymmetric information. Using descriptive analytical methods grounded in prior case-based and empirical studies, the research identifies systematic mechanisms through which contractors and public agencies respond to risk, uncertainty, and institutional constraints. The findings highlight that cost escalation and delays are not merely technical failures but rational responses to misaligned incentives and weak enforcement environments. The article contributes theoretically by bridging construction management research with advanced procurement economics, and practically by outlining policy-relevant insights for designing contracts, managing inflation risk, and improving procurement performance in large-scale road infrastructure. The study concludes that durable reform requires aligning incentives across the project lifecycle, strengthening institutional capacity, and recognizing the endogenous nature of opportunism within public procurement systems.

This article examines 60Out’s Godzilla vs. Kong escape room as a representative case of designing complex narrative–interactive environments in which spectacle, safety, and player engagement are fused into a single managerial system. The relevance of the study is defined by the contemporary evolution of the location-based entertainment industry, which is shifting from traditional game formats to comprehensive immersive spaces that must simultaneously deliver sensory impact, cognitive absorption, and operational reliability. The novelty lies in conceptualizing a model of coordinated orchestration, implemented through the COGS system, which functions as a practical instantiation of the Experience Manager from interactive narrative theory. This mechanism demonstrates the feasibility of integrating three traditionally separate domains—show control, the cognitive dynamics of flow, and engineering safety—into a unified, coherent control architecture. The principal findings indicate that the success of Godzilla vs. Kong derives not only from high-technology components, but from their systemic coordination: synchronization of multimedia to produce spectacle, nonlinear design and adaptive hints to sustain flow, and embedded monitoring and intervention capabilities to ensure safety. It is shown that such integration yields durable business outcomes (a 25% increase in bookings) and establishes the basis for a new disciplinary paradigm in which experience design is understood as the governance of interdependent parameters of risk, engagement, and dramaturgy. The article will be useful to researchers in interactive media, entertainment-industry professionals, and designers of themed environments.

Background: The proliferation of AI-enabled microservices in enterprise environments has necessitated robust strategies for dynamic scaling. While Platform-as-a-Service (PaaS) offerings provide inherent scalability, they often suffer from "cold-start" latency and unpredictable cost implications during varying workloads, such as refinery turnarounds or large-scale data processing events.

Methods: This study introduces a Resilient Scaling Orchestrator (RSO) that integrates Ansible-based automation with formal process algebraic models to optimize end-to-end dynamic scaling. We employ a hybrid methodology that combines theoretical formal component modeling to predict system states with practical chaos engineering experiments to validate resilience. The approach leverages Ansible playbooks to pre-warm instances based on predictive heuristics, mitigating cold-start latency.

Results: Experimental validation using industry-standard microservices benchmarks demonstrates that the proposed RSO reduces cold-start latency by approximately 40% compared to reactive Azure PaaS autoscaling. Furthermore, the integration of formal verification ensures that 99.9% of scaling operations maintain transactional integrity even under induced chaos scenarios.

Conclusion: The findings suggest that combining infrastructure-as-code tools with formal mathematical modeling provides a superior framework for managing the cost-performance trade-off in cloud-native AI applications.

Abstract: In recent years, the global semiconductor supply chain has been subject to multiple, interwoven shocks—ranging from pandemic-induced disruptions, macroeconomic upheavals, geopolitical tensions, to climate‑related risks. This paper offers a comprehensive, theory-driven examination of systemic vulnerabilities in semiconductor supply networks and explores the strategic role of reshoring and resilience metrics in reconfiguring these networks for greater stability and autonomy. Drawing on a multidisciplinary synthesis of supply chain resilience theory, macroeconomic analyses, industry reports, and policy research, we construct a detailed conceptual framework that integrates resilience measurement, disruption impact channels, and strategic adaptations such as reshoring and diversification. The analysis reveals that while traditional supply‑chain resilience metrics (e.g., recovery time, robustness, flexibility) remain essential (Behzadi et al., 2020; Campuzano & Mula, 2011), they are insufficient alone in the semiconductor context: the long lead times, high capital intensity, and geopolitical concentration necessitate additional dimensions—sovereign autonomy, climate‑risk exposure, and macroeconomic elasticity. We show how reshoring initiatives, especially in high‑value segments like GPU manufacturing, can enhance strategic autonomy and buffer macroeconomic vulnerabilities (Lulla, 2025; PwC, 2025). However, reshoring presents trade‑offs: elevated costs, potential innovation slowdowns, and environmental externalities. The paper concludes by offering a set of refined resilience metrics tailored for the semiconductor industry, and a policy‑oriented roadmap for firms and governments to promote a more resilient, adaptive, and autonomous semiconductor supply ecosystem.

The article proposes an original MLOps maturity model specifically oriented toward retail organizations. The relevance of the study stems from the fact that, despite active investment in machine learning, many retailers face difficulties with scaling, ensuring reliability, and assessing the return on investment of AI initiatives. The presented model serves as a roadmap for the phased and systematic development of MLOps practices. The scientific novelty lies in the domain adaptation of general MLOps principles to the retail context and, critically, in establishing clear and measurable criteria for transitions between five maturity levels. The paper analyzes existing universal maturity models. The five levels, from chaotic to optimized, are described through the lens of four key dimensions: technology and data, ML development, deployment and operations, governance and people. Particular emphasis is placed on the development of concrete checklists that make it possible to verify readiness to transition to the next level. The purpose of the study is to provide retail companies with a tool for self-assessment and strategic planning to build their MLOps capabilities. To achieve this goal, methods of analysis of existing models, synthesis, and domain adaptation are used. In conclusion, it is emphasized that a high level of MLOps maturity is primarily a strategic rather than a purely technical task. The material is addressed to CDOs, CIOs, heads of Data Science, and MLOps engineers in retail.

The paper surveys recent advances in generative pipelines that produce and optimize personalized advertising creatives across image and poster formats. The study synthesizes evidence on constraint ingestion, layout-aware rendering, retrieval-assisted staging, human-feedback inspection, CTR-oriented reward conditioning, and serving-time selection/ranking. Particular attention is paid to how knowledge-augmented vision-language adapters improve brand/price text handling; how poster systems encode hierarchy for legibility; how retrieval narrows the feasible space before diffusion; how inspector-driven feedback reduces unusable variants while correlating with live engagement; and how joint or parallel ranking architectures preserve creative diversity without latency penalties. The goal is to develop an operational blueprint for customer acquisition that reduces idea-to-launch cycles while maintaining brand safety and persuasive clarity. Methods employed include a comparative synthesis of ten recent studies, a structured content analysis, and a concept mapping of failure modes, evaluation metrics, and optimization objectives. The findings consolidate an end-to-end stack that aligns offline screening with online lift and reallocates human effort from repetitive triage to governance and experiment design.

Background: The built environment is responsible for substantial material throughput, waste generation, and lifecycle environmental impacts. Scholarship across materials science, construction management, policy, and industrial ecology emphasizes reuse, recycling, and design-for-disassembly as central strategies for decoupling building-sector growth from resource depletion (Jawahir & Bradley, 2016; Jones & Comfort, 2018). Yet implementation remains fragmented by technical, economic, regulatory, and social barriers (Hart et al., 2019; Hjaltadóttir & Hild, 2021).

Objectives: This article constructs an integrative, publication-ready conceptual and methodological framework—grounded exclusively in the provided literature—that synthesizes material-level interventions (e.g., wood-plastic composites, recycled timber), component- and building-level reuse strategies (e.g., whole-house deconstruction, load-bearing component reuse), procurement and policy levers (e.g., circular procurement), and enabling digital-technological tools (e.g., BIM for demolition planning). It aims to reconcile disparate empirical findings into a coherent research agenda and actionable decision framework for practitioners, policymakers, and researchers.

Methods: A critical integrative review approach was used, combining thematic synthesis, cross-case comparative analysis, and systems-mapping. Source materials included experimental materials research, case studies of deconstruction and reuse, life-cycle and economic feasibility analyses, policy and procurement scholarship, and technological studies on digital tools for demolition and waste estimation (Keskisaari & Kärki, 2018; Bouslamti et al., 2012; Zaman et al., 2018; Cheng & Ma, 2013). Each source was interrogated for contribution to technical feasibility, economic viability, regulatory impediments, stakeholder dynamics, and design implications. Claims are triangulated across multiple sources, and where tensions exist, alternative interpretations are explored.

Results: The synthesis identifies four mutually reinforcing domains necessary for scalable circularity in the built environment: (1) material innovation and substitution pathways (e.g., wood-plastic composites employing industrial wastes); (2) building- and component-level recovery systems (e.g., systematic deconstruction and reuse of load-bearing elements); (3) institutional and market mechanisms (e.g., circular procurement and cost-competitive reuse supply chains); and (4) digital and process enablers (e.g., BIM-based waste estimation and mapping of material flows). Critical bottlenecks identified include uncertain cost attribution for secondary materials, quality and performance variability in reclaimed materials, regulatory ambiguity over reused structural elements, and information asymmetries inhibiting reuse markets (Yeung et al., 2017; Sigrid Nordby, 2019; Serwanja & Sheidaei, 2016).

Conclusions: Transitioning to circular built environments requires coordinated interventions across technical, institutional, and informational axes. Design for Reuse, integrated procurement strategies, and digital traceability constitute a combined pathway to reduce lifecycle impacts while maintaining safety and cost-effectiveness. Research priorities include standardized performance metrics for reclaimed materials, procurement models that internalize circularity benefits, and scalable logistics models for deconstruction and material redistribution. The article closes by proposing a detailed research and policy agenda that operationalizes the integrated framework introduced.

Originality/Value: This work pioneers the integration of an event-driven data processing methodology with BIM and the COBie standard specifically for the critical domain of large-scale port infrastructure maintenance, offering a pathway toward a fully realized Digital Twin of port assets.

To ensure performance, safety, and efficiency, thermal management is key to the operation of electric vehicles (EVs) as they continue to scale varying climates, charging behaviors, and duty cycles. This paper describes a path to thermal intelligence which leverages publicly available datasets. Some of these datasets include drive profiles from NREL Fleet DNA, climate data from NOAA GHCN, battery aging data from NASA and MIT, and workplace charging behaviors from ACN-Data. The paper also draws upon open-source simulator or learning tools such as PyBaMM, FASTSim, and pythermalcomfort. Using a combination of physics and machine learning, we obtain a 54% reduction in root mean square error (RMSE) for peak battery temperature predictions based on a physics-only baseline. The smart system utilizes physical and uses machine learning to predict cabin HVAC energy use, given different comfort constraints (PMV/PPD). During experimentations in urban commutes and last-mile delivery, we find that cabin HVAC range reductions can exceed 10% in extreme climates; as a countermeasure, we piloted comfort-aware setpoint relaxations as well as charging-aware pre-conditioning the night before. In the case of charging-aware pre-conditioning, by using real-world timestamps for the charging events, we reduced the starting battery temperature by 6.8°C while simultaneously increasing passenger comfort by 85%. All of this was done without an increase in onboard energy consumption. We believe this work provides for the construction of open thermal intelligence pipelines to maintain safety, efficiency, and comfort for future software-defined Electric vehicle and fleet platforms.

The democratization of investing remains a critical global challenge, particularly for populations excluded from traditional wealth-building tools due to insufficient capital, limited literacy, or systemic barriers. Recent advances in cloud computing and artificial intelligence (AI) offer an unprecedented opportunity to bridge this gap at scale. This paper investigates the end-to-end design, implementation, and evaluation of AI-powered cloud platforms for micro-investment and wealth building. We explore design patterns and architectural paradigms that leverage modular, scalable cloud infrastructure to support a new generation of micro-investment tools. Emphasis is placed on AI-driven recommendation systems, explainability, and the integration of regulatory and security requirements. The system architecture is described in detail, including data ingestion, AI model pipelines, portfolio management, compliance layers, and user interfaces. Implementation frameworks, performance and fairness evaluation criteria, and deployment challenges are presented. The paper concludes with an honest assessment of current technical, ethical, and regulatory limitations, as well as recommendations for the responsible advancement of AI-powered micro-investment platforms.

Explainable AI (XAI) is critical in domains like credit scoring where model decisions must be transparent and accountable. This survey paper compares three local explanation techniques—SHAP, LIME, and ensemble Hybrid approach that integrates both. We evaluate these methods on consistency, variability, and suitability for regulatory environments. Emphasis is placed on use in credit risk modeling, with insights drawn from both literature and practical evaluation.

This paper investigates the transformative effects of autonomous AI systems on human learning and the dissemination of knowledge. It presents a framework for developing self-evolving knowledge solutions that integrate autonomous individuals with adaptive AI networks. By employing continuous feedback loops and dynamic interactions, these systems facilitate a perpetual flow of knowledge, thereby enhancing both individual and collective intelligence. The study highlights the key mechanisms through which AI supports personalized learning experiences and accelerates the evolution of knowledge. It also addresses challenges related to autonomy, scalability, and ethical considerations. The proposed model aims to bridge the gap between human cognition and machine intelligence, fostering a collaborative ecosystem for lifelong learning. This work contributes to the emerging field of AI-driven knowledge management and educational innovation.

Sharing of Cyber Threat Intelligence (CTI) has turned out to be an indispensable pillar of the modern cybersecurity landscape, it is enabling organizations to defend against the evolving threats. In this white paper, we will discuss the strategies to optimize the sharing of threat intelligence across multiple security platforms in the enterprise and community context. We will observe the current standards and practices, like Structured Threat Information eXpression (STIX) and trusted Automated Exchange of Indicator Information (TAXII) protocols, and also examine the role of these standards in integrating the Threat Intelligence Platforms (TIPs) with Security Information and Event Management (SIEM) systems. We will observe the impact of threat intelligence exchange through real-world case studies and how the cybersecurity attacks are mitigated, along with the challenges that are encountered (e.g., technical integration gaps, data overload, trust and privacy issues). We will also discuss the limitations in the current approaches, which include the inconsistent adoption of the standards, there is a prevalence of indicators with low context, and siloed systems that impede the information flow. The landscape of the emerging solutions, the future directions will be explored, machine learning prioritized to reduce the false positives, a decentralized sharing architecture by leveraging blockchain and federated learning for privacy, and also trust frameworks to incentivize collaboration. Through addressing the present challenges and leveraging the advanced technologies, organizations will be able to create a unified and effective threat intelligence sharing ecosystem that will strengthen the collective cyber defense.

Natural Language Processing (NLP) is a key enabler of conversational user interfaces for mobile chatbots and voice assistants, which are used more and more for smart applications such as customer support, personal assistance, and home automation. But deploying NLP models on mobile devices is challenging because these platforms are resource-constrained with limited processing power, memory, and battery life. In this paper, we discuss some of the most important NLP methods like tokenization, text categorization, and entity recognition, which are needed for mobile voice assistants and chatbots. We also discuss how there is a trade-off between local inference, where models are executed on the device, and cloud inference, which provides greater model capabilities at the cost of latency and privacy. Methods to improve NLP models for mobile devices, such as model compression, low-power designs, and hybrid solutions, are explored in great detail. Then, speech recognition integration with NLP in voice assistants is also explored with regard to challenges like real-time processing, privacy, and noise management. We conclude by defining future directions and challenges and highlighting the importance of scalable, energy-efficient, and privacy-preserving NLP systems for mobile devices.

With the rapid adoption of cloud computing, remote collaboration, and digital transformation, organizations face increasing risks from insider threats and data theft. Among these, “outlookers”—malicious insiders, compromised employees, or external adversaries leveraging legitimate access—pose a particularly stealthy and dangerous challenge. Unlike traditional intruders, outlookers exploit trusted credentials to exfiltrate sensitive data while evading perimeter-based defenses and rule-driven detection systems. This paper systematically reviews Artificial Intelligence (AI) and Machine Learning (ML) approaches for identifying and mitigating outlooker activities through continuous monitoring, anomaly detection, and behavioral analytics. Frameworks such as the Insider Threat Kill Chain, Zero-Trust Security Model, and Cybersecurity Maturity Model (CMM) are examined to contextualize AI’s role in strengthening organizational resilience. Case studies from enterprise and government deployments demonstrate that AI-enabled insider threat detection can reduce exfiltration risks by 35–45% while lowering false positives by 20–30%. However, challenges persist in ensuring privacy protection, explainability, and adversarial robustness. The findings underscore that AI-driven solutions represent a critical frontier in safeguarding intellectual property, customer trust, and national security against sophisticated insider threats.

The Inflation Reduction Act (IRA) of 2022 aims to curb soaring pharmaceutical costs by mandating price negotiation for Qualifying Single Source Drugs (QSSDs) that have been on the market for 7 (small molecules) or 11 (biologics) years and lack generic competition. However, this time-based metric introduces a critical systemic vulnerability: Product Hopping (PH). PH is an established anticompetitive tactic wherein manufacturers introduce minor, non-therapeutic reformulations (a New Drug Application (NDA) or Biologics License Application (BLA)) solely to reset the negotiation clock, thereby extending their monopoly. Current reactive regulatory frameworks are insufficient to counteract these complex, data-driven manipulative strategies. This theoretical paper proposes an expert-level conceptual framework for an Artificial Intelligence (AI)-powered Regulatory Surveillance Architecture (RSA) within the Centers for Medicare & Medicaid Services (CMS). This RSA leverages predictive analytics, Natural Language Processing (NLP), and anomaly detection across multi-modal data streams—including patent filings, clinical trial documents, and market data—to quantify the economic and therapeutic rationale underlying reformulation, yielding a Probabilistic Intent Score (PIS). Central to the framework is the mandatory implementation of Explainable AI (XAI) to ensure that regulatory interventions, particularly those triggering high-stakes negotiation, are transparent, auditable, and legally defensible, meeting rigorous standards of administrative due process and governance.

This paper presents an AI-driven framework for real-time monitoring, prediction, and mitigation of urban air pollution in Delhi, India. The proposed system integrates drone-based air quality sensing, IoT-enabled data collection, and AI predictive analytics to forecast pollution levels and recommend proactive interventions. By combining drone data, IoT sensors, and meteorological information, deep learning models forecast pollution spikes and optimize mitigation measures. The system offers a scalable, replicable model for proactive pollution management across global cities.

The exponential growth of large-scale Artificial Intelligence (AI) models, or "Red AI" , has led to a 300,000-fold increase in computational demand since 2012 , raising significant environmental and sustainability concerns. While the high carbon cost of model training (e.g., GPT-3's estimated 550 metric tons of CO2e) is well-documented, this focus obscures the dominant environmental burden: model inference, which can account for up to 90% of a model's total lifecycle energy consumption. A critical research gap exists in the unified analysis of carbon cost versus performance metrics across this entire AI lifecycle. Furthermore, the field lacks a standardized, comprehensive framework for Green AI reporting, hampering transparent and verifiable comparisons. This paper addresses this gap through a systematic review and quantitative synthesis of Green AI. We systematically categorize and evaluate three pillars of technical optimization: (1) model compression, (2) hardware-aware AI, and (3) low-power inference techniques. This analysis reveals that high-level architectural choices—such as using general-purpose generative models for discriminative tasks—are orders of magnitude (e.g., 14.6x to 30x) less efficient than task-specific models. We also highlight a "measurement crisis," where common reporting tools like CodeCarbon underestimate true energy consumption by 20-40% compared to ground-truth measurements. We conclude by proposing a comprehensive, lifecycle-based Green AI reporting framework, designed to integrate with existing GHG and ISO standards. This framework mandates unified cost-performance metrics (e.g., CO2e/ inference / performance-unit) to enable transparent, verifiable, and-informed decision-making for sustainable AI development.